Download Report

INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.3 Issue.4, Pg.: 74-79
April 2015
INTERNATIONAL JOURNAL OF
RESEARCH IN COMPUTER
APPLICATIONS AND ROBOTICS
ISSN 2320-7345
IMPLEMENTATION OF SRAM ARRAY BY
USING MULTIBIT FLIP-FLOP
K. Ravi Teja Kumar1, K. Ramesh2
1
2
M.TECH (VLSI DESIGN) Deportment of ECE SVNE, Tirupathi, [email protected]
ASSISTANT PROFESSOR Deportment of ECE SVNE, Tirupathi, [email protected]
Abstract: - Memory elements play a vital role on Digital World. In memory devices the most important
factor is power consumption. Because the power consumption of the memory device increases means, the
device reliability and life time is reduced. The basic memory elements of designer considerations are Latch and
Flip-flop. In this paper we design SRAM using arrays of flip-flops and we analyze the design of Single-bit Flipflop (SBFF i.e., 1bit) and made performance comparison over the Multi-bit Flip-flop (MBFF i.e., 2bit, 4bit, 8bit,
16bit, and 32bit). While designing the memory by using SBFF means it consumes more power. To get
maximum reduction in power an algorithm has been proposed in which Single-bit flip flops are replaced with
maximum possible Multi-bit Flip-flop without affecting the performance of the original circuit. This paper
analyzes the timing performance of both SBFF and MBFF in Xilinx Virtex-5 family (XC5VLX50). These
results in favour of Multi-Bit Flip-flop as reduction of power and area.
Keywords: Low power, Flip-flop, Latch, Merging, SBFF, MBFF, SRAM.
I. Introduction
In the past decades, the main concerns of VLSI designers were performance, area, cost and reliability. Power
consumption was mostly of only secondary importance to other things. However, this has begun to change and,
with major priority, power consumption is given comparable importance to area and speed. Area optimization
means to optimize (reduce) the space of a digital design, such as that of an integrated circuit, while preserving
the functionality [1]. This is done in both front-end and back-end of design. In front-end design, proper
description of simplified Boolean expression and removing unused states will lead to minimize the
gate/transistor utilization. Partition, Floor planning, Placement, and routing are perform in back-end of the
design.
Power consumption can be divided into two aspects, such as Dynamic power consumption and Static power
consumption. Dynamic power is referred to as the power is consumed by a device, when it is actively switching
from one state to another. Dynamic power consists of switching power, consumed while charging and
discharging the loads on a device, and internal power (also referred to as short circuit power), consumed internal
to the device while it is changing state. Static power is referred to as the power consumed by a device not related
to state changes (also referred to as leakage power). Generally the main concern with leakage power is when the
device is in its inactive state, as all the power consumed in this state is considered “wasted” power.
Designers have developed various low power management techniques to reduce power consumption, such as
power gating creating multisupply voltage designs, clock gating, dynamic voltage/frequency scaling, and
minimizing clock network. These techniques make use of some form of sleep operation. Placing power gating
K. Ravi Teja Kumar & K. Ramesh
Page 74
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.3 Issue.4, Pg.: 74-79
April 2015
structures is a well-known technique for reducing power leakage during standby mode, while maintaining high
speeds in active mode.
II. SBFF and MBFF
In this section, we will introduce single-bit and multi-bit flip-flop conception[2]. Before that, we will review
single-bit flip-flop (1-bit), it has single input is called the “DATA” input. D Flip-flop is one of the most
commonly used Flip-flops. For a Positive-Edge-Triggered D Flip-flop, its output Q follows input D only at
every L to H transition of CLOCK, otherwise, Q keeps unchanged. Figure 1 shows the RTL schematic of Single
bit flip-flop (1-bit).
Fig. 1 RTL schematic of Single bit flip-flop (1-bit)
From the timing diagram in fig 2 it is clear that the output Q changes only at the positive edge. At each positive
edge the output Q becomes equal to the input D at that instant and this value of Q is held until the next positive
edge.
Fig. 2 Timing diagram for SBFF
The concept of merging some 1-bit flip-flops into a multi-bit flip-flop is applied to reduce dynamic clock power
and decrease the total flip-flop area in a synchronous design[4]. A single-bit flip-flop has two latches (Master
latch and slave latch).The latches need “Clk” and “Clk’ ” signal to perform operations, such as Fig. 3 shows.
The multi-bit Flip-flops are mostly viewed as low power design technique, MBFFs with larger bit numbers as
possible to gain more clock power saving but larger bit number may lead to severe crosstalk’s due to close
interconnecting wires. The single bit Flip-flop and Merging of two 1-bit flip-flops shown in the below figures.
Fig. 3 merging two single bit flip-flops into one 2-bit flip-flop
Multi-bit flip-flop (2-bit, 4-bit, 16-bit, 32-bit etc....) has two or more inputs. For example consider 4-bit flipflop, it has 4 inputs and 4 outputs shown in the below figure.
K. Ravi Teja Kumar & K. Ramesh
Page 75
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.3 Issue.4, Pg.: 74-79
April 2015
Fig. 4 RTL schematic of multi bit flip-flop (4-bit)
Multi-bit Flip-flop which takes multiple data input and results in multiple data output. The working of multi-bit
Flip-flop is same as single-bit Flip-flop, whenever the clock gets active state Flip-flop latches all input to output.
For inactive state the Flip-flop holds the data. The multi-bit Flip-flop corresponding waveform is given in fig 5.
Fig. 5 Timing wave for MBFF
The MBFF has many advantages. That is
i. Area and delay is reduced.
ii. Number of inverters is reduced in clock sinks. So total power consumption is reduced.
III. SRAM
The data storage in SRAM depends on the d.c source. If the d.c is removed means, the data can be erased.
SRAM means Static random-access memory, it is a type of semiconductor memory that uses bistable latching
circuitry to store each bit. The term static differentiates it from dynamic RAM (DRAM) which must be
periodically refreshed. By using the Multi-Bit Flip-Flops the SRAM is designed.
SRAM 4k-bit has own input and output lines and has control signals, WRITE and PHI_b. SRAM fully operates
in static mode. Therefore, no clock or refreshment is required. Testing for 4K-bit SRAM flows along the
functional blocks address decoders, SRAM cell and multiplexers [3].
Fig. 6 4k-bit SRAM Functional block diagram
K. Ravi Teja Kumar & K. Ramesh
Page 76
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
A
Address Input
D_IN
Data Input
D_OUT
Data Output
WRITE
Write Command Input
PHI_b
Bit line Pre-charge Command Input
Vol.3 Issue.4, Pg.: 74-79
April 2015
Table I. Pin description
SRAM cell has four cases for its operation; read “1” or “0”, write “1” or “0” and all of these cases were tested
using Xilinx Virtex-5 family (XC5VLX50). Because SRAM cell requires a sort of tuned timing in input signals,
each of PHI_b, WL, WRITE and DATAIN was set up to meet this requirement. By using address lines the flipflop is select to write or read the data. The operation of the SRAM is, when r=0, the write operation is done.
When r=1, the read operation is performed.
IV. Simulation Results
The analysis of SRAM designed using MBFF is targeted and verified on Xilinx FPGA of family Spartan-3E.
The constraints taken to considerations are number of flip-flops used, clock buffer count and period analysis as
Gate delay and net delay. The number of Flip-flops in SRAM using SBFF is about half a number in SRAM
using MBFF. The strategies are to minimizing the power dissipation.
The Proposed system is simulated and verified using Verilog HDL in Xlinix ISE 10.1i for the target device
xc3s500e-5fg320. The below simulation results show the outputs of the SBFF and MBFF.
Fig. 7 Simulation result for single bit Flip-flop (1-bit)
Fig. 8 Simulation result for multi bit Flip-flop (4-bit)
K. Ravi Teja Kumar & K. Ramesh
Page 77
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.3 Issue.4, Pg.: 74-79
April 2015
The synthesis result of SBFF(1-bit) and MBFF(2-bit, 4-bit, 16-bit and 32-bit) device utilization, cell usage and
number of IOs shown in the below Table II, Table III and Table IV respectively.
Bit type
No. Of
No. of
IOs
Bonded IOBs
1-bit
3
3
2-bit
5
5
4-bit
22
20
16-bit
66
64
32-bit
130
128
Table II. Device utilization report
Bit type
IO Buffers
IBUF
OBUF
1-bit
2
1
1
2-bit
4
2
2
4-bit
20
10
10
8-bit
64
32
32
16-bit
128
64
64
Table III. Cell usage
The below Table represents the comparison of proposed method with existing method in terms of IOs
Bit type
IOs
1-bit
3
2-bit
5
4-bit
22
8-bit
66
16-bit
130
Table IV. Comparison of SBFF and MBFF IOs
V. Conclusion
In this paper, we presented Multi-Bit Flip-flops in combination with SRAM array to reduce the power
consumption. The Single-Bit Flip-Flop consumes more power because of inverters in clock sinks. To reduce
this power consumption MBFF is designed. In here Xilinx Virtex-5 family (XC5VLX50) tool is used to design
the SRAM using MBFF. The SBFF SRAM dissipates the power as 324.16 mW and MBFF SRAM dissipates
the power as 323.93 mW. So, 0.23 mW power dissipation is reduced. The Simulation Results and Power
Analysis Report indicate that multi-bit flip-flop in combination with SRAM array is very effective and efficient
method in lower-power designs.
K. Ravi Teja Kumar & K. Ramesh
Page 78
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.3 Issue.4, Pg.: 74-79
April 2015
REFERENCES
[1] G.Prakash, K. Sathishkumar, S. Saravanan, “Achieveing reduced area by multi-bit Flip Flop design”
International Conference on Computer Communication and informatics, Jan. 04-06,2013.
[2] Challa Sankaramma, K D Mohana Sundharam, “Optimization of area through multi bit flip flop design
using FPGA” IRF International Conference, 14th December 2014.
[3] Joon-sung yang/ Gahngsoo Moon, Dec. 9, 2005 “32k-bit sleepy SRAM”. Benton H. Calhoun, Frank A.
Honore, Anantha P.Chandrakasan-“A Leakage reduction methodology for distributed MTCMOS”.
[4] LI Xia Yu, JIA Song, LIU LiMin, WANG Yuan and ZHANG GangGang , “ Design of Novel, Semitransparent flip-flops for high speed and low power application”, science china Press and Springerverlag Berlin Heidelberg ,2012.
K. Ravi Teja Kumar: He is currently Studying M.Tech-VLSI Sree Vidyanikethan Engineering College,
Tirupati. His areas of interest are Digital System Design, VLSI Design.
K. Ramesh: He is currently working as an Assistant Professor in ECE department of Sree Vidyanikethan
Engineering College, Tirupati. He has completed M.Tech in VLSI Design, in NIT Jamshedpur. His research
areas of interest are Digital Design, VLSI Signal Processing.
K. Ravi Teja Kumar & K. Ramesh
Page 79